Neurosurgery 97:481–488, 2025
Automated quantitative CT analysis using normalized volumetric CT density (nv-CTD) enhances prediction of basal ganglia hematoma expansion, especially when combined with the spot sign. nv-CTD offers high sensitivity for ruling out expansion and enables improved risk stratification for early intervention in intracerebral hemorrhage.
• Automated quantitative CT analysis (nv-CTD) predicts hematoma expansion (HE) in basal ganglia intracerebral hemorrhage (ICH), improving risk stratification for early surgical intervention.
• nv-CTD is calculated as mean ICH CT density divided by surrounding parenchyma density, providing a normalized measure of hemorrhage acuity.
• Lower nv-CTD (<2.3) is highly sensitive (96%) for predicting HE, while the spot sign is highly specific (95%); using both improves diagnostic accuracy (AUC 0.80 vs 0.68 for spot sign alone).
• nv-CTD alone performs similarly to the spot sign for HE prediction and can be used when CTA is unavailable, especially to rule out HE.
• Automated computer vision segmentation enables consistent, rapid, and reproducible feature extraction, overcoming limitations of manual CT interpretation.
• The study included 108 patients and used a custom-trained neural network for image analysis, excluding those with thalamic ICH, high IVH burden, or small hemorrhages.
• Limitations include retrospective design, modest sample size, and narrow inclusion criteria, which may limit generalizability.
• Volumetric and quantitative imaging analysis can augment clinical decision-making for basal ganglia ICH management.
You must be logged in to post a comment.